History · dating the past
A beam from a ruined house has no date written on it. But its rings do: a pattern of wide and thin years that matches one other piece of wood in the world, and that match can name the year to the season.
In trees from places with real seasons, each growing season adds one new layer of wood under the bark. That layer is a ring. In a good year, with plenty of water and warmth, the ring comes out wide. In a drought or a cold snap it comes out thin.
So a trunk is a bar chart of the weather, oldest year in the middle and newest at the edge. And here is the useful part: trees growing in the same region feel the same weather. A drought that squeezed one pine also squeezed its neighbours, and a pine cut a hundred kilometres away usually shows the same thin ring in the same year.
Counting rings gives you an age. It does not give you a calendar date, because you don't know which year the tree stopped growing. The calendar date comes from a trick called crossdating: slide the unknown pattern along a record whose years are already known, until the wiggles line up.
Below is a made-up record of 160 years, 1700 to 1859, drawn as one bar per ring. A mystery beam has been cut from somewhere inside that period. Drag it, or use the slider, and watch the match score. It runs from 1 for a perfect copy down to 0 for no relationship.
Drag the teal beam along the record. The red band marks which years of the master it is sitting on. The line at the bottom shows the score for every possible position.
The ring widths in this bench are invented for the demo. Real chronologies are measured from cores and beams, but they behave the same way.
Switch the beam to 10 rings. Now a few other spots in the record score nearly as well as the right one. That is not a bug in the method. Ten rings is a short sentence, and short sentences can match by accident. A 25-ring pattern is much harder to fake, and at 50 rings a wrong position falls well away from the true one.
This is why dendrochronologists like long samples, and why they check that a match holds over several trees instead of trusting one. Two independent pieces of wood that both agree on the same year are very hard to explain by luck.
The beam doesn't need to be old. It only needs enough rings to be unlike every other stretch of the record.
The astronomer A. E. Douglass started on tree rings in the 1890s, when he was working at Lowell Observatory in Arizona. Decades later, in 1919, he compared wood from two big ruins in the American Southwest, Aztec Ruin and Pueblo Bonito, and found that Pueblo Bonito had been built 40 to 45 years before Aztec. He had no calendar date yet, but he could already say which came first and by how much.
The calendar date needed the same trick we used above, repeated across generations. Start with living trees, whose outer ring is this year. Match their inner rings to the outer rings of old stumps and beams cut from nearby, then match those to older beams, and so on backwards. The record Douglass built from living pines and ruin timbers had a gap for a while, a stretch of undated "floating" ruins. A charred beam from Show Low, Arizona, catalogued as HH-39, is the one traditionally credited with bridging it. Accounts of exactly when differ, but by 1929 the Southwest record ran back continuously to about the year 700. In his December 1929 report Douglass put Aztec Ruin's timbers at 1111 to 1120.
Other regions did the same. The longest chain on record, built from oaks and pines in central Europe and anchored at the Hohenheim laboratory, now reaches back 12,460 years. California's bristlecone pines give a continuous record of about 8,500 years, and the oldest living bristlecones are roughly 4,900 years old.
Crossdating is precise, but it has rules. The wood needs enough rings, and the reference record has to come from the same species and region, because a wet year in Arizona says nothing about a beam from Bohemia. A match dates the ring, not the day someone built with it. To say when the tree was cut, you need the outermost ring, ideally with bark still attached. A beam trimmed square, or reused from an older house, tells you less.
The payoff goes beyond archaeology. Tree rings help date old violins and wooden panels for paintings, and they provide the year-by-year checks used to calibrate radiocarbon dating. The international radiocarbon calibration curve relies on them for roughly the last 13,900 years.
A tree has no idea it's writing a calendar. It just grows a bit more in a good year, and that is enough.
Facts drawn from the University of Arizona Laboratory of Tree-Ring Research history pages, SAPIENS and standard dendrochronology references. Dates for the Hohenheim and bristlecone chronologies are as currently published; the demo record is invented.